融入注意力机制的多模特征机械臂抓取位姿检测
作者:
作者单位:

西南科技大学 信息工程学院

作者简介:

通讯作者:

中图分类号:

TP242

基金项目:

国防科工局项目([2019]1276), 国家自然科学基金项目(12175187),西南科技大学博士基金项目(19zx7123)


Multi-modal feature robotic arm grasping pose detection with attention mechanism
Author:
Affiliation:

School of Information Engineering, Southwest University of Science & Engineering

Fund Project:

The Project of State Administration of Science, Technology and Industry for National Defence , PRC ([2019]1276), The National Natural Science Foundation of China (12175187)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    针对机械臂抓取检测任务中对未知物体抓取位姿检测精度低、 耗时长等问题, 提出一种融入注意力机制多模特征抓取位姿检测网络. 首先设计多模态特征融合模块, 在融合多模态特征同时对其赋权加强; 接着针对较浅层残差网络提取重点特征能力较弱的问题, 引入卷积注意力模块, 进一步提升网络特征提取能力; 最后, 通过全连接层对提取特征直接进行回归拟合, 得到最优抓取检测位姿. 实验结果表明, 在 Cornell 公开抓取数据集上, 本文算法的图像拆分检测精度为 98.9 %, 对象拆分检测精度为 98.7 %, 检测速度为 51 FPS, 对10类物体的100次真实抓取实验中, 成功率为 95 %.

    Abstract:

    To address the problems of low accuracy and time consuming detection of unknown object grasping pose in robotic arm grasping detection task, a Multi-modal feature grasping pose detection network with attention mechanism is proposed. The first is to design a Multi-modal feature fusion module to fuse the Multi-modal features and enhance their weighting; then, to address the problem that the shallow residual network is weak in extracting key features, a convolutional attention module is introduced to further improve the feature extraction ability of the network; finally, the optimal grasp detection pose is obtained by direct regression fitting of the extracted features through the fully connected layer. The experimental results show that the detection accuracy of image splitting and object splitting on Cornell Grasp dataset is 98.9 % and 98.7 %$ respectively, and the detection speed is 51 FPS. The success rate is 95 % for 100 real-world grabs of 10 types of objects.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2022-05-10
  • 最后修改日期:2023-02-06
  • 录用日期:2022-09-06
  • 在线发布日期: 2022-09-17
  • 出版日期: